Intraperitoneal (ip) cancer occurs frequently in gastrointestinal and gynecologic malignancies. The restricted volume of the peritoneal cavity and the limited egress from this space make possible selective drug exposure through ip administration. The therapeutic potential of antitumor agents may be enhanced by ip administration to take advantage of a steep dose-response. The prerequisites for successful ip therapy include a safe and reliable drug delivery system, adequate distribution of the therapeutic agent so that all tumor bearing sites are exposed and, the use of appropriate drug regimens. Totally implanted silastic catheter/port systems will be tested for their ability to facilitate ip chemotherapy and fluid sampling without dressing changes or infection. Prior pharmacology studies of ip chemotherapy have utilized repeated lavages or short infusions of large volumes of chemotherapy containing fluid on the assumption that such is necessary to ensure adequate distribution within the abdomen. The volumes administered present logistical difficulties and usually require hospitalization. A major objective of this proposal is to determine what fluid volume provides optimal ip drug/fluid distribution by comparing conventional contrast assisted CT scanning and nuclide tomographic scanning techniques. Subsequently, osmotically active high molecular weight substances such as Dextran 70 and albumin will be administered ip in an attempt to retain a satisfactory ip volume of fluid over several days and thus facilitate outpatient protracted infusion regimes utilizing small portable infusion pumps. The effect of such high molecular weight agents on pharmacokinetic parameters will be determined by comparison with previous studies which utilized large volume infusions. Agents suitable for ip use must have a high total body clearance and a relatively low clearance from the peritoneal cavity to provide a selective regional advantage. New pharmacologically rational agents for ip administration will be studied including fluorodeoxyuridine (FUDR), dichloromethotrexate (DCMTX), and bischlorethylnitrosourea (BCNU). Concurrent ip and plasma levels of FUDR and DCMTX will be measured with protracted infusion to determine regional selectivity. The toxicity and regional selectivity of bolus ip BCNU will be similarly determined. Ultimately drug combinations will be tested in pilot efficacy studies using the delivery mechanisms and regimens derived in the studies above.